Skip to content
Tree Search Molecular Dynamics Simulation
Jupyter Notebook Python
Branch: master
Clone or download
Latest commit 1a71b7b Aug 25, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
.ipynb_checkpoints first commit Mar 4, 2019
init first commit Mar 4, 2019
result first commit Mar 4, 2019
LICENSE Initial commit Mar 4, 2019
README.md
analysis.ipynb first commit Mar 4, 2019
initialize.py first commit Mar 4, 2019
pacs_md.py first commit Mar 4, 2019
tsmd.py first commit Mar 4, 2019
viz_util.py fix bug Mar 7, 2019

README.md

TSMD

Python implementations of TS-MD* and PaCS-MD**. Our implementations use GROMACS command for all of the manipulation about MD simulation. The required version is GROMACS 2016.5.

* K. Shin, D. P. Tran, K. Takemura, A. Kitao, K. Terayama, and K. Tsuda, "Enhancing biomolecular sampling with reinforcement learning: tree search molecular dynamics simulation method," ACS Omega, 2019. [DOI: 10.1021/acsomega.9b01480]

** R. Harada and A. Kitao, "Parallel cascade selection molecular dynamics (PaCS-MD) to generate conformational transition pathway," J. Chem. Phys., 139, 07B611 1. (2013)

USAGE

equivration

Before running TS-MD or PaCS-MD, we have to do adding ion, energy minimization, nvt equilibration and npt equilibration. We need the .mdp files to do each manipulation of such preparations. The .mdp files here are intended only for use with this chignolin(1uao) folding sample. You have to set the parameters for each of your tasks.

initialize.py() 

Requrements

Both TS-MD and PaCS-MD program require the files shown below.

  • reactant structure (.gro)
  • checkpoint file after the equilibration (.cpt)
  • topology file (.top)
  • target structure (.gro)
  • mdp file for short md (.mdp)

TS-MD

  • Options
    • -r : Reactant structure file (.gro).
    • -t : Target structure file (.gro).
    • -top : Topology file (.top).
    • -s : Step size. default is 1000.
    • -c : A parameter of ucb socre. Default is 0.05.
    • -cn : Resume from the previous state. Default is 0.
    • -ntmpi : The number of mpi.
    • -ntomp : The number of open MP
    • -del : If you want to delete the intermediate trajectory files, set this to 1. default is 0.
    • -thresh : A thereshold (Å) that decide whether one structure is similar to others.
    • -alp : A penalty parameter.
    • -ctype : Choose the way how to calclate UCB. (normal, adaptive, adaptive2)

If you have the reactant structure file as reactant.gro, target structure file as target.gro, and topology file as topol.top your command is

python pats_md.py -r reactant -t target -top topol

If TS-MD doesn't reach the enough RMSD, the intermediate state is preserve as vars.pikcle automatically. You can resume the process by setting the -cn to 1.

PaCS-MD

python pacs_md.py -r reactant -t target -top topl
  • Options
    • -r : Reactant structure file (.gro).
    • -t : Target structure file (.gro).
    • -top : Topology file (.top).
    • -s : Step size. Default is 1000.
    • -k : The number of parallelization. Default is 5
    • -cn : Resume from the previous state. Default is 0.
    • -ntmpi : The number of mpi.
    • -ntomp : The number of open MP.
    • -del : If you want to delete the intermediate trajectory files, set this to 1. default is 0.
You can’t perform that action at this time.